Beyond curse of dimensionality with transition matrices

We implement the column-by-column way to the model with structural changes.

for

for the first group of period of time

for the second group of period of time

.

.

.

for the ns-th group of period of time

together with different parameters across groups.

We set non-aggregate model, aggregate model, and general model.

We perform the same modifications the to get v1 and indMs without expanding utility

matrix as yesterday but we use collocation method to get the convergence today.

Notice that it will take more memory to run because the part of collocation method

takes up lots of memory.